> certain frequencies for them anyway, so we just use that and set them up If I am not mistaken, Allwinner recommends > - The basic clocks (OSC24M, OSC32K, AHB1, APB1, APB2, PERIPH0) are > driven clock system using the SCPI framework. > So as mentioned several times before, I am looking into a more firmware This is admittedly also true for the old clock > clocks, since they would have to copy quite verbatim the Linux > - It makes it quite hard for any other DT user (*BSD, U-Boot) to use the Simple-panel binding is a similar example. A recent thread adding display timings in DT to the I'm guessing this was nacked by device tree maintainers and/or clk > a generic sunxi-ng user, which explains the needed clocks in the DT > us back to something like the old sunxi clock scheme, which is > data and looks like it should really be presented in DT - which brings > - This code file does not actually contain any code, instead it's just > SoC, even if they share 95% of the clocks (as H3 and A64 do, for instance). > the future), it ultimately requires an explicit code drop for every new Although that makes the DT pretty easy (and avoids breaking it > - sunxi-ng requires a complicated SoC specific source file in the > Basically I see those issues with the new clocks:
> - We upstream some basic DT first, possibly omitting any controversial > we may move this into a possible arisc firmware. > - ARM Trusted Firmware provides the SCPI implementation - for now, later > specific clocks (apart from the simple-gates). > allwinner,simple-gates and fixed clocks. > - We consider using an SCPI based clock system for the A64, alongside > As this became quite a long read, here a TL DR: > clock system, so I use that opportunity here -) > As I think I never officially expressed my concerns about the new sunxi > latest local branch), I will comment on this later. > I think I have some small fixes to the DT (have to compare against my > In general this looks good to me - apart from the sunxi-ng clock usage. > Thanks very much for looking into this and compiling this series! > already has a decent amount of devices supported: uart, i2c, rsb, etc. > Some WiP stuff have been removed, such as the MMC part, but this serie > first patches are just meant to rework slightly the H3 code, before > This uses the current H3 clock code, as both are really similar. > Here is the previous A64 patches made by Andre, reworked to use On Mon, at 9:43 AM, Andrà Przywara wrote: In reply to: Andre Przywara: "Re: arm64: Allwinner A64 support based on sunxi-ng".Previous message: Benjamin Tissoires: "Re: power: surface3_power: MSHW0011 rev-eng implementation".
0 kommentar(er)
